Within 25 years, the US population aged 65 and over will double in size to 80 million bringing, with it an epidemic of aging-related cognitive decline, from normal cognitive aging to neurodegenerative disorders including Alzheimer?s Disease (AD). These conditions impair quality of life and functional status, impose an enormous burden on individuals, their families, the healthcare system, and require elucidation of mechanisms and development of new treatments to prevent or at least slow their progression. The use of plant-based food and drink for health purposes has a long and well-documented history. Cocoa beans contain the flavanol epicatechin, an anti-oxidant with beneficial effects on blood pressure, endothelium-dependent vasomotor function, platelet reactivity, insulin sensitivity, vascular inflammation, and circulating progenitor cells. Importantly, flavanols have neuroprotective effects, suppressing oxidative stress and inflammation and promoting neurogenesis, neuronal survival and synaptic plasticity, all of which are relevant to the pathophysiology of neurodegenerative disorders like AD, amyotrophic lateral sclerosis (ALS), and Parkinson?s Disease (PD). Evidence from humans, non-human primates, and rodents points to a role for the hippocampus and its subregions in aging-related neurodegenerative disorders including AD. We recently reported that dietary intake of cocoa flavanols increased hippocampal function, measured as fMRI cerebral blood volume (CBV) and a pilot mediation analysis showed that cocoa flavanols led to a decrease in the sentinel pro-inflammatory mediator HMGB1, an activator of the innate immune system. In turn, this decrease in HMGB1 was linked to improved hippocampal function. Recent evidence implicates HMGB1 in cognitive decline and impairment. HMGB1 binds to Toll-like receptor 4 (TLR4), triggering the production of pro-inflammatory cytokines including TNFa via NFkB-dependent pathways. In rodent models of endotoxemia and surgical trauma, HMGB1 mediated hippocampal-dependent memory impairment similar to that seen in septic patients, an effect eliminated by neutralizing HMGB1. These data mechanistically link HMGB1 to neurodegenerative impairment, suggesting its potential as a therapeutic target, consistent with evidence that amplified systemic inflammation is associated with a variety of age-related pathologies including Alzheimer?s Disease. They strongly support our major hypothesis that cocoa flavanols improve hippocampal function by their effects on neuroinflammation, specifically HMGB1, via a TLR4-NFkB-TNFa signaling pathway. We propose to test this model in a randomized controlled trial of 146 participants, age 50-69, receiving high or low daily cocoa flavanol for 12 weeks. Such a trial has potential for significant clinical impact.